Probability, Probability-Based, and Nonprobability Surveys in Psychiatric Epidemiological Research.

Area of Science:

• Psychiatric epidemiology
• Population health surveillance
• Survey methodology

Background:

• Global rise in mental health challenges necessitates robust surveillance.
• Traditional probability sampling for psychiatric surveys faces financial and operational hurdles.
• Online panels offer cost-effectiveness and speed but introduce potential biases.

Purpose of the Study:

• To establish a standard for utilizing both nonprobability and probability-based online panel surveys in psychiatric epidemiology.
• To propose a hybrid design balancing the strengths of different survey methodologies.

Main Methods:

• Recommends integrating online panel surveys with small, parallel high-quality probability samples.
• Utilizes probability samples as surrogates to understand online panel participant differences.
• Includes measures strongly associated with outcomes and panel membership in both surveys.
• Employs statistical methods to blend results from both sample types.

Main Results:

• Hybrid designs mitigate biases inherent in online panels.
• They also overcome operational difficulties associated with high-quality probability samples.
• This approach enhances the practicality and validity of large-scale epidemiological studies.

Conclusions:

• Hybrid designs represent the minimally acceptable standard for psychiatric epidemiological surveys of the household population.
• This methodology ensures more reliable population-level insights for evidence-based policy.
• Alternative designs are noted for rapid, less expensive exploratory analyses.